That would be evidenced pretty simply by whether the players and GM were engaging or avoiding the system I would imagine. It's been my experience that normally good GMs make bad systems work largely by ignoring them.
Personally I'm not so interested how it handles in the hands of a bad GM, a bad GM can wreck any game, good or bad, so I don't see much point in writing toward the lowest common denominator as it often disempowers better GMs from doing what they do best. So rather than attempting to GM-proof the system, I would prefer they empower GMs to run their games how they wish while simultaneous giving guidance and advice on how to be better, rather than attempting to protect the players from them.
On the other hand, there isn't a game in existence than can make playing under a bad DM fun, amiright?
That's great advice for a published game or a publicity game (which DDXP really was), but a lousy way to run a playtest.
The designers need feedback from the groups, especially the DMs about what doesn't work well in practice and the only true way to tell is to run it. Hopefully, the open playtest is not the first chance of such feedback from an audience outside the designer circle since small groups tend to develop assumptions/blind spots quickly. I hope they've been running a larger closed playtest to look for areas in the rules that are generally ignored to improve play. Those are the areas that will need the most attention to make a good game.
Otherwise you'll likely make a game that needs to be ignored in play.
It's not really advice, so much as a statement that I would prefer that bad GMs not be the metric around which the game is calibrated. I agree wholeheartedly that the designers should get as much and widely varied feedback as they can. I'm in no way suggesting that it only be given over to a select group of highly trained GMs only to be taken for a test drive under the most favorable conditions, I was merely reacting to the title of the thread.